[Moims-dai] Telecon Notes?

D or C Sawyer Sawyer at acm.org
Fri Jul 27 01:44:09 UTC 2018


Dear John, David, Mark, and Steve,

It seems I’ve not been able to clarify my concerns regarding the rigidity with which the OAIS RM information modeling is being viewed, at least by some, in the thread below, but I believe the following scenario will highlight some differences and clarify my views.

I’m attending a conference on preservation and I’m giving a presentation on my Archive implementation.  I state that my implementation follows OAIS RM precepts and that I will be using OAIS RM terminology and concepts to describe the implementation.  I describe how the implementation is addressing the high level bulleted requirements of  OAIS RM Section 3.  I describe my AIP implementation as consisting of a Tar file with pointers to the relevant information, including Representation Information for the Content Data Object and PDI consisting of Provenance, Context, Reference and Access Rights Information.

I then note, as others have noted in past years, that one of the great features of the OAIS RM is that it includes both digital and physical information at the most fundamental level, the Information Object, as my Archive has both types of materials.  Not only do I have a variety of digital materials, but I also have a number of historical manuscripts in hardcopy form that need preservation.  In addition, a number of my digital Content Data Objects are quite old and their format information (Representation Information) is stored in hardcopy form on microfiche. I’m able to describe all this using the OAIS RM terms and concepts.

Of course my AIPs for the historical manuscripts look a bit different from those for the digital Content Data Objects.  They have a pointer to a physical storage locations for each manuscript.  I describe how access to these manuscripts is controlled, and how the environment of these storage areas is controlled to help reduce degradation of the manuscripts over the long term (Fixity).  I also note that the Designated Community for these manuscripts is able to read and understand some, for which this is the preservation objective, without any additional supporting information, and for others they only need to be made visible to experience and no further understanding is part of the preservation objective.  Therefore none of their AIPs include pointers to Representation Information.  Further, none of them have pointers to Access Rights information as this does not exist.  In fact, some of my digital Content Information objects do not have Access Rights information either.

As I wind up my presentation, I feel that I’ve been able to communicate my implementation to the audience quite well and I ask for questions.  Immediately a hand shoots up from the back of the room.  He stands up and says “Mr. Sawyer, I think there is a problem here.  You started your presentation by saying that you implemented an OAIS archive.  You described your AIP implementations for the manuscripts as not having pointers to Representation Information and I understand you don’t need any.  However in the OAIS RM conformance section it says for an implementation to be considered an OAIS, among other criteria that you clearly fulfill, it must support the information model of Section 2 and this model shows an AIP has Representation Information.  Therefore it seems to me that some of your AIPs are not in conformance with the Section 2 AIP information model and thus your archive is not a true OAIS. I also note that you are missing some Access Rights information, although this is not mentioned explicitly in Section 2.”

I respond that as far as I’m concerned, I do support the Section 2 AIP information model.  I’ve used all the OAIS RM concepts and implemented them in a manner that makes sense for the particular nature of the materials involved and I’ve used these concepts in communicating my implementation to this audience.  I point out that the main purpose of the OAIS RM is to promote just this type of communication. Another major purpose is to highlight the types of information and their relationships to aid in long term preservation.  My presentation has done that as well.  That the manuscripts do not need Representation Information is an important point in carrying out their long term preservation according to their associated preservation objectives.  When I read ‘support’ in Section 2, I take it to mean that I should apply the AIP information model concepts to my implementation in a manner that is logically consistent with the AIP components and their relationships.  I do not take it to mean that I should apply the concepts in a way that is rigid and not consistent with the nature of the materials and is in fact extraneous to the preservation of those materials.

I continued, saying that I believe my implementation is a quintessential example of an OAIS.  Unless the OAIS RM makes clear that only a rigid implementation of all or certain features of the information modeling is needed to be called an OAIS, I will continue to call my implementation a full OAIS.  However I do not expect to see this nor do I think it would be wise.  Regardless, my implementation will stand as what I believe it is and that is an outstanding example of the full application of OAIS RM concepts and relationships.

Thanks for your attention.

Cheers-

Don

PS  I’ll respond to individual comments below, if requested, but I believe the scenario above should make my position clear and may prompt different questions/comments.


> On Jul 24, 2018, at 10:06 AM, Hughes, John S (398B) <john.s.hughes at jpl.nasa.gov> wrote:
> 
> Hi all,
>  
> Sorry that I have been responding more regularly on this topic but there have been so many other fun things to do.
>  
> I tend to agree with John’s view and concerns. However I will try to address the issue from a different point of view.
>  
> I believe that everyone would agree that a physical book in the context of an OAIS archive is an information object. I can accept that there is no need for Representation Information in the strict sense, i.e., information that maps the Data Object into more meaningful concepts, for example a data format.
>  
> However definitional information is require, for example “the data object is a hardcopy book”. I have accepted Mark’s suggestion that the definitional information can be recast as representation for the purpose of maintaining the integrity of the OAIS Information Object.  This is a rather useful expedient otherwise the PDS4 information model collapses. Also defining something is in some ways a mapping a thing to a more meaningful concept. However strictly speaking, defining something is the first step in providing meaning.
>  
> So my position is that some information, call it representation information or definitional information is required for every data object (CDO).
>  
> Thanks,
> Steve
>  
>   <>
> From: MOIMS-DAI [mailto:moims-dai-bounces at mailman.ccsds.org] On Behalf Of John Garrett
> Sent: Monday, July 23, 2018 10:03 PM
> To: 'MOIMS-Data Archive Interoperability' <moims-dai at mailman.ccsds.org>
> Subject: Re: [Moims-dai] Telecon Notes?
>  
> Hi Don,
>  
> Hope things  are going well with you and the family.  I hope you’re making progress with the renovations. David is finishing his.  Time for me to start some more.
>  
> Now down to work.
>  
> I have to say that I still think that if a physical book is the CDO for an AIP then Rep Info is still required.  The Rep Info may be minimal and may be as simple as saying “The CDO is a physical book and can be interpreted by simply viewing/reading it.
> Since an AIP is a logical construct, if the whole repository is composed of physical books, then the implementation could be that the above statement could logically be considered as the Rep Info for all the AIPs. 
>  
> I also don’t see why the made up sentence is a different meaning for Rep Info.  It seems to me that the made up sentence is explaining how to make sense of the CDO as much as any other Rep Info. 
>  
> Conceptually, I don’t understand how a implementation that logically does not have any Rep Info can also be considered to logically have Rep Info.
>  
> Who decides if a CDO, physical or digital, requires Rep Info if it is not always required?
> Does a physical book require Rep Info if it is in a foreign language?
>  
>  
> If everyone else agrees that Rep Info is not required, we can make that change.
> But if we do mean to make that change, we certainly need to change all the figures (and any related text) that show one-to-one relationships between the CDO and Rep Info in the OAIS standard.
>  
> Peace and joy,
> -JOhn
>  
>  
> From: MOIMS-DAI [mailto:moims-dai-bounces at mailman.ccsds.org <mailto:moims-dai-bounces at mailman.ccsds.org>] On Behalf Of D or C Sawyer
> Sent: Sunday, July 22, 2018 9:28 AM
> To: MOIMS DAI List <moims-dai at mailman.ccsds.org <mailto:moims-dai at mailman.ccsds.org>>
> Subject: Re: [Moims-dai] Telecon Notes?
>  
> Dear all,
>  
> I would like to return to this point in the discussion on David’s proposal as I think David has made some valuable points that may get lost.  I have not heard anyone disagree that, under some circumstances, a physical book does not need RepInfo.  As an example, I believe an Archive implemented primarily for the preservation of physical books (Content Data Objects) understandable to the associated Designated Community would be fully justified in not conjuring up information, to be called RepInfo, just to meet someone’s concept of a best practice or a mandatory requirement for some type of certification.  There may, or may not, eventually be a requirement for the addition of real RepInfo to such CDOs.  This is all about practical implementations.  
>  
> In contrast, the OAIS RM is a conceptual model and its AIP is the key unit of information that is being preserved.  The AIP is defined to include RepInfo with the CDO, forming Content Information, as this Content Information entity is of central importance to the preservation of digital information.  I see no reason to alter this Content Information definition and I don’t believe anyone is proposing to do so. At the same time I have no problem recognizing that some implementations, such as that given in the Archive example above, may have an AIP whose Content Information is the CDO alone. This simply reflects local, practical, reality. The example highlights the possible differences between a concept and local reality.  In my opinion, when there is good justification, such differences should not be of major concern. David’s proposal for the OAIS RM regarding implicit and explicit RepInfo is designed to acknowledge this reality.  While I had some objection to the ‘implicit’ part of his proposed wordingl, I firmly support the attempt and believe some appropriate words can be found. It should not take a lot of discussion to find such words if there is agreement on the objective.
>  
> When the OAIS RM uses the phrase ‘Representation Information’, it should mean the same concept everywhere in the document.  It should not also stand for some made up sentence just to satisfy a conceptual information category.  Therefore I think John’s second paragraph below is properly understood as an implementation view where he thinks an actual AIP should always have a RepInfo category.  One could consider this a possible best practice, but personally I would not go that far.  Discussion of this particular view  makes sense in the context of ISO 16363.  As for the OAIS RM, again I support David’s attempt to clarify that some AIP implementations may not need explicit RepInfo as it has been defined in the OAIS RM.   
>  
> Cheers-
> Don
>  
>  
> 
> On Jul 17, 2018, at 4:11 AM, John Garrett <garrett at his.com <mailto:garrett at his.com>> wrote:
>  
> Hi,
>  
> I’m also uncomfortable with the updated conformance.  I also feel that it would be a significant change and I think it would open up a door to keep expanding the conformance requirements.  I would prefer to try to update Section 2 a bit and keep the Conformance section the same.
>  
> Also I’m not on board with having original data objects in an AIP that do not require RepInfo.  I think if we have a physical book in local language, we should still require RepInfo as we form the AIP.
>  
> Peace and joy,
> -JOhn
>  
>  
> From: MOIMS-DAI [mailto:moims-dai-bounces at mailman.ccsds.org <mailto:moims-dai-bounces at mailman.ccsds.org>] On Behalf Of D or C Sawyer
> Sent: Monday, July 16, 2018 11:36 PM
> To: MOIMS DAI List <moims-dai at mailman.ccsds.org <mailto:moims-dai at mailman.ccsds.org>>
> Subject: Re: [Moims-dai] Telecon Notes?
>  
> Dear All,
>  
> The proposal for section 1.4 (conformance) is to include section 4.2 information modeling for conformance. This is a considerable increase in requirement which originally we did not feel comfortable requiring.  This is not backward compatible so it may be controversial in the larger community.
>  
> The proposal for section 4.2.1.3 states that an explicit or implicit lack of Representation Information can be regarded as a sufficient expression of meeting a Representation Information requirement. There is a bit of a problem with ‘implicit’ as it could be used as justification for no Representation Information when in fact it was needed.  Perhaps the ‘implicit’ should be dropped?
>  
> However this kind of statement on Representation Information is also true for selected components of PDI (which can also be recursive), depending on the judgement of the Archive in support of the associated Designated Community. I believe some appropriate text is needed for PDI as well, particularly if 4.2 is now to be part of conformance.
>  
> Once again I will be unable to participate in the Tuesday DAI telecon as I’ll be working with my contractor.
>  
> Cheers-
> Don
>  
> 
> 
> 
> On Jul 16, 2018, at 8:00 AM, David Giaretta <david at giaretta.org <mailto:david at giaretta.org>> wrote:
>  
> I have added a suggested change  - see http://review.oais.info/show_bug.cgi?id=236 <http://review.oais.info/show_bug.cgi?id=236>
>  
> Also, in the resolution of http://review.oais.info/show_bug.cgi?id=35 <http://review.oais.info/show_bug.cgi?id=35> there was an action in comment 4 to create a related suggested change which I have done in http://review.oais.info/show_bug.cgi?id=237 <http://review.oais.info/show_bug.cgi?id=237>
>  
> ..David
>  
> From: MOIMS-DAI <moims-dai-bounces at mailman.ccsds.org <mailto:moims-dai-bounces at mailman.ccsds.org>> On Behalf Of David Giaretta
> Sent: 15 July 2018 19:24
> To: 'MOIMS-Data Archive Interoperability' <moims-dai at mailman.ccsds.org <mailto:moims-dai at mailman.ccsds.org>>
> Subject: Re: [Moims-dai] Telecon Notes?
>  
> Hi Don and others in MOIMS-DAI
>  
> It looks as if we can reach a conclusion – see below for some general suggestions. I will try to suggest something specific on the review site.
>  
> Regards
>  
> ..David
>  
> -----Original Message-----
> From: MOIMS-DAI <moims-dai-bounces at mailman.ccsds.org <mailto:moims-dai-bounces at mailman.ccsds.org>> On Behalf Of D or C Sawyer
> Sent: 13 July 2018 14:18
> To: MOIMS DAI List <moims-dai at mailman.ccsds.org <mailto:moims-dai at mailman.ccsds.org>>
> Subject: Re: [Moims-dai] Telecon Notes?
>  
> Hi David, et al.,
>  
> I mostly agree with your comments, but see below.
>  
> > On Jul 9, 2018, at 7:12 AM, David Giaretta <david at giaretta.org <mailto:david at giaretta.org>> wrote:
> > 
> > Hi all
> > 
> > I missed the call on which these notes report, but I must make a few comments which I am sure I would have made in the meeting.
> > 
> > In summary Don's arguments about the Information Model seem to derive from examples of some specific Data Objects, without explicit consideration of the Designated Community, which I argue is misleading, and can be addressed by adding text to the standard to address such "edge cases", as discussed next.
>  
> During the subject telecon it was mentioned that a CDO with no needed Representation Information depended on the Designated Community. The point of the discussion was to note that this is a real example. It could be thought of as an ‘edge case’, as you say, or perhaps the modeling and text should be improved to bring it into compliance.
>  
> Yes we definitely should add text to explain this, perhaps in the discussion in section 4.2.1.3 Representation Information.
>  
> > 
> > The argument about whether or not a Data Object needs Representation Information in order to be understood is misleading in that, as I responded in a previous email, it depends on the Designated Community. In other words, if the Data Object is a piece of paper written in English, and members of the Designated Community can read and understand English then there is no need for an explicit English dictionary because the Designated Community's Knowledge Base includes that. 
> > 
> > However, if the Designated Community were to be defined as readers of Chinese then Representation Information such as an English Dictionary would be needed. So, if one claims that this piece of paper does not need Representation Information then that is in the context of a Repository's definition of a specific Designated Community. 
> > 
> > So in general modelling terms the OAIS Information Model is right - a Data Object needs Representation Information in order to be used/understood. However there may be some specific instances where, for a specific Designated Community, explicit Representation Information is not needed.
> > 
> > In modelling terms, I would regard cases such as English written on paper to be read by English readers as an "edge case" and to be consistent in such edge cases one should say that the Representation Information is the statement that "for the specific Designated Community no additional explicit Representation Information is currently needed". As far as I can see this would make the model complete and consistent. We could add some wording to the discussion of Representation Information and Designated Community to make this clear.
> > 
> > In terms of the idea of OAIS being only a "communications framework" - yes it is that, especially the Functional Model, but it also provides something more, namely guidance of what one needs in order to preserve digitally encoded information, and OAIS lays these out in terms of conformance. Thus, the repository needs an AIP with all its components; it needs to undertake all the Mandatory Responsibilities. 
>  
> Yes on the Mandatory Responsibilities, but on the ‘AIP with all its components’ there is considerable ambiguity.  The information modeling conformance only addressed the Information Object and the Information Package as given in Section 2.2. The Information Package is broken into SIPs, AIPs, and DIPs,  It is clear that SIPs and DIPs may not contain all the Representation Information or PDI.  The implication (not stated) is that the AIP does contain all the Representation Information and PDI.  Certainly it does conceptually.  However the conformance section (1.4) addresses an actual Archive.  The example above raises the conformance issue for an Archive that only preserves hardcopy documents, with no Representation Information, yet still fully understandable to its Designated Community, Maybe it also finds it doesn’t need all the PDI.  Can it still claim to be OAIS conformant? I would say ‘yes’ as long as it is making a reasonable argument for not having all the components of the section 2.2 information models.  Note also that none of the modeling in section 4.2 is required for OAIS RM conformance.  Given that OAIS RM conformance has sometime been a topic of concern to others as well as to this working group, I recommend it be clarified.
>  
> I agree – we should make it clear that 2.2 and 4.2 are both included in the Mandatory Responsibilities and also in the Conformance section. In section 2.2 we say “While formal modeling of information is provided in section 4, some key concepts are provided in this subsection” but this should be expanded also.
>  
> > 
> > I realise that to some readers all this may seem like a very abstruse, uninteresting discussion. However, it IS important because there will be times when repository staff will come across puzzling cases and we need to make sure, as far as humanly possible, that OAIS provides useful, correct and consistent guidance when read carefully.
>  
> Yes, we should want the OAIS RM to be as clear and unambiguous as we can make it.
>  
> Yes I think we have relied on some implicit connections that should be made explicit.
>  
> Cheers-
> Don
>  
>  
> > 
> > Hope that helps
> > 
> > ..David
> > 
> > 
> > 
> > -----Original Message-----
> > From: MOIMS-DAI <moims-dai-bounces at mailman.ccsds.org <mailto:moims-dai-bounces at mailman.ccsds.org>> On Behalf Of John Garrett
> > Sent: 07 July 2018 05:55
> > To: 'MOIMS-Data Archive Interoperability' <moims-dai at mailman.ccsds.org <mailto:moims-dai at mailman.ccsds.org>>
> > Subject: [Moims-dai] FW: Telecon Notes?
> > 
> > 
> > 
> > -----Original Message-----
> > From: Don Sawyer [mailto:topcottageguys at gmail.com <mailto:topcottageguys at gmail.com>] On Behalf Of D or C Sawyer
> > Sent: Monday, July 2, 2018 8:18 AM
> > To: John Garrett <garrett at his.com <mailto:garrett at his.com>>
> > Cc: Hughes, John S (398B) <John.S.Hughes at jpl.nasa.gov <mailto:John.S.Hughes at jpl.nasa.gov>>; Mark Conrad <Mark.Conrad at nara.gov <mailto:Mark.Conrad at nara.gov>>
> > Subject: Re: Telecon Notes?
> > 
> > Hi John,
> > 
> > I’ve made a few comments to your comments.  I think our differing views are pretty clear.  If you want to add some more comments, or not, I have no objections to your sending this out to the wider group.
> > 
> > Cheers-
> > Don
> > 
> > 
> > _______________________________________________
> > MOIMS-DAI mailing list
> > MOIMS-DAI at mailman.ccsds.org <mailto:MOIMS-DAI at mailman.ccsds.org>
> > https://mailman.ccsds.org/cgi-bin/mailman/listinfo/moims-dai <https://mailman.ccsds.org/cgi-bin/mailman/listinfo/moims-dai>
>  
> _______________________________________________
> MOIMS-DAI mailing list
> MOIMS-DAI at mailman.ccsds.org <mailto:MOIMS-DAI at mailman.ccsds.org>
> https://mailman.ccsds.org/cgi-bin/mailman/listinfo/moims-dai <https://mailman.ccsds.org/cgi-bin/mailman/listinfo/moims-dai>
> _______________________________________________
> MOIMS-DAI mailing list
> MOIMS-DAI at mailman.ccsds.org <mailto:MOIMS-DAI at mailman.ccsds.org>
> https://mailman.ccsds.org/cgi-bin/mailman/listinfo/moims-dai <https://mailman.ccsds.org/cgi-bin/mailman/listinfo/moims-dai>
>  
> _______________________________________________
> MOIMS-DAI mailing list
> MOIMS-DAI at mailman.ccsds.org <mailto:MOIMS-DAI at mailman.ccsds.org>
> https://mailman.ccsds.org/cgi-bin/mailman/listinfo/moims-dai <https://mailman.ccsds.org/cgi-bin/mailman/listinfo/moims-dai>
>  
> _______________________________________________
> MOIMS-DAI mailing list
> MOIMS-DAI at mailman.ccsds.org
> https://mailman.ccsds.org/cgi-bin/mailman/listinfo/moims-dai

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.ccsds.org/pipermail/moims-dai/attachments/20180726/8e53c9ec/attachment.html>


More information about the MOIMS-DAI mailing list